Study on the assessment method of regional disaster based on the change of cholorophyll-a concentration in seawater

Di Wu*#, Yu Zhang*#, Hao Wang§, Mengxing Huang*#, Wenlong Feng*#, Rouru Chen*# *State Key Laboratory Marine Resource Utilization in South Sea, Haikou, 570228, China #College of Information Science and Technology, University, Haikou, 570228, China §Big Data Lab, Dept. of ICT and Natural Sciences, Norwegian University of Science and Technology, Aalesund, 6009, Norway

Abstract—Due to global climate change, the frequency and index and damage potential index to make the impact analysis uncertainty of typhoon become more and more prominent. When of typhoon for rubber tree plantations. However, these works a typhoon disaster occurs, it’s an imminent problem to estimate mainly focus on the route prediction or the degree of overall the disaster zone and help ships to move to safer waters. Existing impact, lacking the assessment of the impact on different studies on typhoon disaster are mainly based on the overall wind regions along the typhoon route. assessment or the route prediction of the typhoon, with much less attention on the detailed impact in different regions along the In this paper, with satellite data, we propose a new method route. In this paper, we propose a new estimate method based on by studying the chlorophyll-a concentration in seawater which the assessment of chlorophyll-a concentration in seawater. We (1) is calculated by chlorophyll-a inversion operation with study Landsat-8 satellite images under typhoon weather and Landsat-8 sensing data. According to the regular change of normal weather in the same area; (2) calculate chlorophyll-a chlorophyll-a concentration in the seawater during typhoon concentration, based on which we (3) build a typhoon disaster weather[5][6], we build a new typhoon disaster model, named model. The experiment shows that our model can assess the Chlorophyll-a Typhoon Disaster Model (CTDM), based on the impact of a typhoon in different regions based on the level of regression model between growth rate of chlorophyll-a chlorophyll-a concentration. concentration, intensity and moving speed, by Shao [7]. In particular, CTDM has the following advantages: Keywords—typhoon disaster assessment; chlorophyll-a; Firstly, it does not need to directly measure the wind speed on typhoon hazard model the sea surface instead it assesses the impact of a typhoon by I. INTRODUCTION the change of chlorophyll-a concentration in the seawater when the typhoon is coming. Secondly, a more detailed assessment According to the analysis of 92 in the South can be performed in different regions, compared to existing China Sea during 1998-2009 by Yang and Tang [1], typhoon is methods. mainly generated by tropical depression above the sea level. When the typhoon is coming, there will be a cold vortex in the II. BACKGROUND upper ocean environment. The ocean thermal structure is not stable because of the cold vortex, so that the cold water below A. Chlorophyll-a Inversion Model the mixed layer is more likely to be brought to the upper layer The principle of chlorophyll-a inversion by the satellite of the sea to cause phytoplankton bloom and increase the remote sensing technology is through the sensor of satellite to chlorophyll-a concentration [1][2]. detect the surface of seawater and obtain relevant hydrological A traditional method to studying typhoon is to use fixed- information. After reflection and scattering on the sea and location buoys, so that the change of the water surface in those spontaneous radiation, the electromagnetic waves emitted by spots can be recorded. This method is inefficient, time- the sensor carry the information on the , consuming, and can monitor only a small area. Along with the sea surface roughness and the concentration of various development of advanced satellite remote sensing technology, substances in seawater back to the sensor. The sensor can using satellites to monitor and study typhoon disaster is more measure the energy of different wavelengths of the efficient and effective due to large observation area, time and electromagnetic wave, according to the analysis of the energy, space synchronization, and continuity, and etc. For example, we can deduce a variety of marine physical quantities such as He and Zhang [3] monitored tropical storm “Sepat” and the concentration of chlorophyll-a in the seawater. typhoon “Fitow” about their development route and its landing In this study, the band ratio method is used to establish the on Southeast China using the Fengyun-3 satellite Liu et al. [4] chlorophyll-a inversion model. This model is as follows: used Landsat-8 data, combined with Hainan Island topographic  Yu Zhang and Mengxing Huang are the co-corresponding authors (Email: [email protected], [email protected].). Where is the concentration of chlorophyll-a; is the atmosphere density of typhoon area; is the wind and are respectively expressed in the near infrared band speed of the circumferential tangential direction. and the red band; and are the coefficients. The Coriolis force is a fictitious force used to explain a B. The Regression Model deflection in the path of a body moving in latitude relative to The correlation between typhoon and chlorophyll-a is the earth when observed from the earth. This force is described complicated and influenced by many factors. It is more as: effective to set up the optimal combination with multiple （ ） independent variables to predict and estimate dependent  variables than to use one independent variable. （ ） The regression model is built with the change rate of is a constant determined by the selected region: chlorophyll-a concentration and the intensity of typhoon and ; is the angular velocity of the rotation of the earth; the speed of typhoon, which comes from the experiment with is the dimension value of the selected area, which is positive in the data from the by Shao [7]. the northern latitude area and negative in the southern latitude area.  The friction force is caused by the rubbing action of the Where is the change rate of chlorophyll-a surface of earth, as follows: concentration; is the intensity of typhoon; is the speed of typhoon; , and are the regression coefficient and （ ） estimated by data; is the random error.  （ ） C. Fujitais Empirical Formula is the friction coefficient; and are the speed of wind; A great deal of experiments confirmed that the longitudinal and are the factor of wind shear. distribution of the of typhoon above the sea surface can meet the requirements of Fujitais empirical III. CHLOROPHYLL-A TYPHOON MODEL formula [8]. Fujitais empirical formula is as follows: Generally, the friction coefficient is estimated by the

 undermined parameters. And we make . Then the （ ） equation (2-4) can be changed as follows: is the atmospheric pressure of typhoon above the

 sea surface; is the standard atmospheric pressure (generally 1000hpa); is the intensity factor of typhoon, which also is We make integration and induction operation between the pressure difference between the environment and the center equation (2-3) and equation (3-1), as follows:

of typhoon; R is the scale factor typhoon; is ,  used to standardize typhoon. D. The Balance Model of Four Forces of Typhoon The typhoon parameters are determined by using the measured data and the data fitting method. The scale factor of Typhoon is a weather system with very strong self- typhoon and the friction factor are got by this way: the scale organization, the structure of typhoon is clear and has a specific factor =21.9 and the friction factor = . The law. Typhoon in the movement is mainly affected by pressure parameters are taken into equation (3-2), then we get the gradient, centrifugal force, Coriolis force and friction force. equation about the wind speed of typhoon as follows: The four forces balance in the ideal condition, as follows:

 

due to the horizontal pressure difference, and it pushes The equation (2-2) is modified and integrated with the the air particles from the high-pressure area to the low-pressure equation (3-3) to obtain the disaster model of wind power in area, as follows: the different and concrete region of typhoon area based on the

 change rate of the concentration of chlorophyll-a. The chlorophyll-a typhoon model is described as:

Р is the atmospheric pressure of the observed place; is the distance from the observed place to typhoon. 

Because of the inertia of an object that makes a circular motion, there is always a tendency for the object to fly along IV. THE MODEL IMPLEMENTATION the tangential direction of the circle. This tendency is described as: In order to build our new disaster model, we select and classify the original satellite data, then according to the

 typhoon path, we find the satellite data of typhoon disaster area.

Fig. 1 is the framework of the data flow that we used to In this experiment, we selected two satellite images have implement new model. the same central dimension (N15.9º; E111.9º) at different times. Fig. 2 is the satellite image of the normal weather on August 31, 2015. Fig. 3 is the satellite image when the typhoon “Mujigae” came on October 2, 2015.

Fig. 2. The image of normal weather Fig. 3. The image of “Mujigae” The information on “Mujigae” is obtained by the South China Sea typhoon network (http://typhoon.hainan.gov.cn/). B. Study Area According to existing research, a typhoon will cause a large area of precipitation. The precipitation over the land of the coast will make the phytoplankton and the soluble organic matter into the coastal seawater, which will cause the increase of chlorophyll-a concentration in the coastal seawater [9]. At the same time, with the increase of population and the Fig. 1. The framework development of industry and agriculture in recent years, the eutrophication of coastal seawater has been aggravated [10]. In this paper, the value of chlorophyll-a concentration in The above interference factors will affect the experimental the seawater of the study area is obtained by satellite data results. Therefore in this study, we chose the study area near preprocessing and chlorophyll-a inversion, and the related the Xisha Island because of less interference factors. The Fig. 4 experiments are based on the ENVI (v5.3) software platform is the detailed assessment area selected according to the route by Esri. The correlation coefficient is obtained by the of typhoon “Mujigae” . simulation on the Matlab software platform. We combine the correlation coefficient and the inversion result of chlorophyll-a In Fig. 4, the central latitude of the study area is N15.9ºand concentration to build the new disaster model, then, the model the central longitude of the research area is E111.9º. We is applied to evaluate the impact in different regions of typhoon selected point A(N16.572º; E112.154º), point B(N16.570º; area. E112.162º), point C(N16.449º; E112.612º), point D(N16.395 The data flow of the framework makes the experimental º; E113.903 º) and point E(N16.387º; E112.915 º) as the process modular, and reduces the repetition rate of the sample points in the study area. experimental procedure. In Fig. 4, the time of the satellite image in the study area is Beijing time on October 2, 2015 at 11 a.m. At the same time, V. EXPERIMENTS AND METHODS the position of typhoon “Mujigae” is shown as the small circle. A. Data The central latitude is N16.713ºand the central longitude is The experimental data is from NASA Landsat-8 satellite. E119.3ºin the small circle. With the continuous movement of The data is in TIFF format, including 11 bands information, typhoon “Mujigae”, we obtained the nearest position of image files, the file of quality evaluation and the metadata of typhoon “Mujigea” to the study area through the path of TXT format. The quality evaluation file mainly includes the typhoon, as the bigger circle (the central latitude is N19.003º information of shooting time, solar elevation, latitude and and the central longitude is E114.201º) in Fig. 4. The time of longitude of the satellite data and so on. Landsat-8 data can be the bigger circle is Beijing time on October 3, 2015 at 4 p.m. used to study in different fields by different band combinations. Therefore, the Landsat-8 data can be used to calculate the C. Experimental Environment and Method inversion of chlorophyll-a concentration in seawater. By means On ENVI (v5.3) software platform, we obtained the level of of radiometric calibration and atmospheric correction for chlorophyll-a concentration in the seawater by satellite Landsat-8 satellite data, the effects of atmospheric interference, inversion. Fig. 5 is the curve of spectrum of seawater before the systematic noise, the posture of sensor and so on can be atmospheric correction. Fig. 6 is the curve of spectrum of removed. seawater after the atmospheric correction. By comparing Fig. 5 and Fig. 6, it can be concluded that satellite images can reflect  The relevant parameters of the sample points are the real situation of the solar radiation after the atmospheric brought into the equation (3-4) the new disaster model correction. to obtain the calculated value “S” of the wind power of typhoon: SA, SB, SC, SD, SE. The visualization, programming and the numerical calculation of the experiment are done in Matlab. The  We made the curve of the typhoon grate “s”, the curve correlation coefficient of the new typhoon disaster model is of the chlorophyll-a concentration “CHL-a” and the obtained after a binary linear regression analysis. curve of the wind power value “S” respectively. The steps of the experimental method are as follows: According to the comparison of the different tendencies of the above three curves, we determined if it is possible  To obtain the typhoon grate “s” of sample points: sa, sb, to assess the typhoon disaster in advance by studying sc, sd, se. the chlorophyll-a concentration of different sample points in seawater when the typhoon is coming.  By the chlorophyll-a inversion operation, to calculate the chlorophyll-a concentration “CHL-a”: CHL-aA, CHL-aB, CHL-aC, CHL-aD, CHL-aE.

Fig. 4. The moving path of typhoon “Mujigae”

Fig. 5. The curve of spectrum of seawater before the atmospheric correction Fig. 6. The curve of spectrum of seawater after the atmospheric correction

mainly concentrated in 0-9 (μ g/L) and is in a reasonable range [11]. VI. RESULTTS AND DISCUSSION A. Chlorophyll-a Concentration Fig. 7 is the distribution histogram of chlorophyll-a concentration in the study area. According to the distribution histogram, we can find the concentration of chlorophyll-a is Matlab. We brought the correlation coefficient and the change rate of chlorophyll-a concentration in the sample points into the new typhoon disaster model, then we got the calculated value “S” of the wind power of typhoon. As shown in the Fig. 9, the value of SA is 33.714(m/s), the value of SB is 33.715(m/s), the value of SC is 33.766(m/s), the value of SD is 33.796(m/s) and the value of SE is 33.795(m/s). And the values of the wind power of the typhoon are conformed to the national standards for tropical cyclones in China, as shown in the TableⅠ. C. The Discussion Between Chlorophyll-a Concentration And Fig. 7. The distribution histogram of chlorophyll-a concentration The Wind Power Value Fig. 10 shows the distribution of chlorophyll-a At the same time, we obtain the value of the sample points concentration of the normal weather on August 31, 2015 in A, B, C, D, and E. As shown in the Fig. 8, the value of point A seawater of the study area. Through Fig. 10, we can see the is 5(μ g/L), the value of point B is 5(μ g/L), the value of point concentration of chlorophyll-a distributed evenly in the normal C is 6(μ g/L), the value of point D is 8(μ g/L) and the value of weather condition, and the main concentration is concentrated point E is 8(μ g/L). on about 4(μ g/L). B. The Wind Speed of Typhoon Fig. 11 shows the distribution of chlorophyll-a The correlation coefficient of the new typhoon disaster concentration, when typhoon “Mujigae” was coming on model is obtained after the binary linear regression analysis in October 2, 2015.

Fig. 8. The chlorophyll-a concentration of sample points Fig. 9. The wind assessment value of sample points

TABLE I. NATIONAL STANDARDS FOR TROPICAL CYCLONES IN CHINA National Standards for Tropical Cyclones in China Grade of Tropical Cyclones The maximum average The maximum wind near wind speed near the center of the center of the bottom(level) the bottom

Tropical depression 10.8~17.1 6~7 Tropical storm 17.2~24.4 8~9 Severe tropical storm 24.5~32.6 10~11 Typhoon 32.7~41.4 12~13 Severe typhoon 41.5~50.9 14~15 Super typhoon 16 or above

The white part of the image is the cloud brought by is 9.99(μ g/L). The place of the highest concentration is near typhoon “Mujigae”. In Fig. 11, we can see the concentration of the area of the blast ring of typhoon “Mujigae”. The chlorophyll-a in the seawater of the study area has a irregular concentration of chlorophyll-a in the rest area is different distribution when typhoon “Mujigae” was approaching. The according to the wind power of typhoon “Mujigae”. chlorophyll-a concentration in seawater of the study area is in 0.11-9.99(μ g/L), and the highest chlorophyll-a concentration By comparing Fig. 10 and Fig. 11, we can draw several Through the above experiments, it can be concluded that conclusions as follows: when typhoon comes, the change of the content of chlorophyll- a in seawater can be used to evaluate the typhoon’s wind force  When a typhoon comes, it will cause the change of in the region. The content of chlorophyll-a in different regions chlorophyll-a concentration in the seawater. of the seawater indicates that the wind force is different in  The concentration of chlorophyll-a in seawater is different regions. When the content of chlorophyll-a in related to the wind power of a typhoon. seawater is higher, the wind force is stronger. The experimental results above, confirmed that our new typhoon  Stronger wind means, higher chlorophyll-a disaster model (3-4) is feasible and effective for the prediction concentration in the same area. of typhoon disaster. The model also provides a new way to We compare the chlorophyll-a concentration with the wind study the regional disaster of typhoon. According to the power of typhoon “Mujigae” of the sample points in Fig. 12, chlorophyll-a concentration in the seawater of the typhoon, the and obtain the relationship between them. The Fig. 13 is the ship can choose the path with slower wind to reduce the loss typhoon level of the sample points. caused by the typhoon disaster. By comparing Fig. 12 and Fig. 13, we can see that the variation tendencies of chlorophyll-a, the wind power and the wind level of the sample points are the same.

Fig.10. The distribution in the normal weather Fig. 11. The distribution in typhoon “Mujigae”

Fig.12. The relationship between CHL-a value and wind power Fig.13. Typhoon level of sample points utilizing the variation of chlorophyll-a in seawater in the  assessment of typhoon disaster. It provides a new way of thinking for further studies of the use of hydrological VII. CONCLUSION information in the typhoon weather for the assessment of The study in this paper presents a new typhoon disaster regional disaster. At the same time, it can also help the ship to model, forming a theoretical basis for the exploration of avoid the loss of life and property during the typhoon weather. However, there are still some limitations in this study, which REFERENCES could be improved in next steps: [1] Yang Xiao-xia, Tang Dan-ling, Location of sea surface temperature cooling induced by typhoon in South China Sea. Journal of Tropical  The research data based on the data of NASA Landsat- Oceanography,2010, 29(4):26-31. 8 satellite, which can be disturbed by the clouds in the [2] Zhou L, Tan Y, Huang L, et al. Phytoplankton growth and typhoon weather. It is suggested that microwave remote microzooplankton grazing in the continental shelf area of northeastern sensing satellite data can be used, for example, the South China Sea after Typhoon Fengshen. Continental Shelf Research, satellite data of GF-3, which has been launched in 2011, 31(16):1663-1671. China. At the same time, it can also measure the [3] He Jieying, Zhang Shengwei. Introduction of MWHTS onboard FY-3C concentration of chlorophyll-a in seawater by means of Satellite and Typhoon Detecting. Remote Sensing Science, 2015(2):17- water buoy or coastal radar microwave detection. 24. Through the fusion of multi-source data, the [4] Liu S, Zhang J, Cai D, et al. Application of Landsat 8 in monitoring of rubber plantation typhoon disaster. Journal of Natural Disasters, 2016. measurement accuracy of chlorophyll-a concentration [5] Liu X, Wang M, Shi W. A study of a Hurricane Katrina-induced in seawater can be improved. phytoplankton bloom using satellite observations and model simulations.  Our experiment only considered part of the South Journal of Geophysical Research Oceans, 2009, 114(C3):819-834. [6] Lin, I. ‐ I. Typhoon ‐ induced phytoplankton blooms and primary China Sea, and further validation of the model should productivity increase in the western North Pacific subtropical ocean. be done in other regions. Journal of Geophysical Research Oceans, 2012, 117(117):3039. [7] Shao Jinchao. The impact of tropical cyclone translation speed and ACKNOWLEDGMENT intensity on phytoplankon chlorophyll-a concentration in the northern This research is supported by the National Natural Science South China Sea. Ocean University, 2015. Foundation of China (Grant #: 61462022), the National Key [8] Wang Yunfeng, Lu Qiang, Wang Bin, Comparative Analysis of Four Technology Support Program (Grant #: 2015BAH55F04, Initialization Schemes in Typhoon Numerical Simulation. Meteorological Science And Technology. 2011 Grant #:2015BAH55F01), Major Science and Technology [9] Davergne T. Offshore and nearshore chlorophyll increases induced by Project of Hainan province (Grant #: ZDKJ2016015), Natural typhoon winds and subsequent terrestrial rainwater runoff. Marine Science Foundation of Hainan province (Grant #: 614232 and Ecology Progress, 2007, 333(1):61-74. Grant#:617062), Scientific Research Staring Foundation of [10] Li Zhao, Song Shuqun, Li caiwen. Distribution of chlorophyll a and its Hainan University (Grant #: kyqd1610). correlation with the formation of hypoxia in the Changjiang River Estuary and its adjacent waters. Marine Sciences, 2016, 40(2):1-10. [11] Luan Hong. Based on Landsat8 the suspended sediment concentration and chlorophyll a concentration of the Pearl River Estuary remote sensing inversion and time and spatial change. Guangdong Ocean University, 2016.